October 2023
·
10 Reads
·
2 Citations
Vaccine
This page lists works of an author who doesn't have a ResearchGate profile or hasn't added the works to their profile yet. It is automatically generated from public (personal) data to further our legitimate goal of comprehensive and accurate scientific recordkeeping. If you are this author and want this page removed, please let us know.
October 2023
·
10 Reads
·
2 Citations
Vaccine
February 2023
·
65 Reads
·
4 Citations
Blood Cancer Journal
Monoclonal gammopathy of undetermined significance (MGUS) is a benign hematological condition with the potential to progress to malignant conditions including multiple myeloma and Waldenstrom macroglobulinemia. Medications that modify progression risk have yet to be identified. To investigate, we leveraged machine-learning and electronic health record (EHR) data to screen for drug repurposing candidates. We extracted clinical and laboratory data from a manually curated MGUS database, containing 16,752 MGUS patients diagnosed from January 1, 2000 through December 31, 2021, prospectively maintained at Mayo Clinic. We merged this with comorbidity and medication data from the EHR. Medications were mapped to 21 drug classes of interest. The XGBoost module was then used to train a primary Cox survival model; sensitivity analyses were also performed limiting the study group to those with non-IgM MGUS and those with M-spikes >0.3 g/dl. The impact of explanatory features was quantified as hazard ratios after generating distributions using bootstrapping. Medication data were available for 12,253 patients; those without medications data were excluded. Our model achieved a good fit of the data with inverse probability of censoring weights concordance index of 0.883. The presence of multivitamins, immunosuppression, non-coronary NSAIDS, proton pump inhibitors, vitamin D supplementation, opioids, statins and beta-blockers were associated with significantly lower hazard ratio for MGUS progression in our primary model; multivitamins and non-coronary NSAIDs remained significant across both sensitivity analyses. This work could inform subsequent prospective studies, or similar studies in other disease states.
November 2022
·
6 Reads
·
1 Citation
Blood
July 2022
·
154 Reads
·
9 Citations
Frontiers in Big Data
Electronic vaccine certificates (EVC) for COVID-19 vaccination are likely to become widespread. Blockchain (BC) is an electronic immutable distributed ledger and is one of the more common proposed EVC platform options. However, the principles of blockchain are not widely understood by public health and medical professionals. We attempt to describe, in an accessible style, how BC works and the potential benefits and drawbacks in its use for EVCs. Our assessment is BC technology is not well suited to be used for EVCs. Overall, blockchain technology is based on two key principles: the use of cryptography, and a distributed immutable ledger in the format of blockchains. While the use of cryptography can provide ease of sharing vaccination records while maintaining privacy, EVCs require some amount of contribution from a centralized authority to confirm vaccine status; this is partly because these authorities are responsible for the distribution and often the administration of the vaccine. Having the data distributed makes the role of a centralized authority less effective. We concluded there are alternative ways to use cryptography outside of a BC that allow a centralized authority to better participate, which seems necessary for an EVC platform to be of practical use.
June 2022
·
76 Reads
·
9 Citations
Mayo Clinic Proceedings Innovations Quality & Outcomes
Objective To assess the generalizability of a clinical machine learning algorithm across multiple emergency departments (EDs). Patients and Methods We obtained data on all ED visits at our health care system’s largest ED from May 5, 2018, to December 31, 2019. We also obtained data from 3 satellite EDs and 1 distant-hub ED from May 1, 2018, to December 31, 2018. A gradient-boosted machine model was trained on pooled data from the included EDs. To prevent the effect of differing training set sizes, the data were randomly downsampled to match those of our smallest ED. A second model was trained on this downsampled, pooled data. The model’s performance was compared using area under the receiver operating characteristic (AUC). Finally, site-specific models were trained and tested across all the sites, and the importance of features was examined to understand the reasons for differing generalizability. Results The training data sets contained 1918-64,161 ED visits. The AUC for the pooled model ranged from 0.84 to 0.94 across the sites; the performance decreased slightly when Ns were downsampled to match those of our smallest ED site. When site-specific models were trained and tested across all the sites, the AUCs ranged more widely from 0.71 to 0.93. Within a single ED site, the performance of the 5 site-specific models was most variable for our largest and smallest EDs. Finally, when the importance of features was examined, several features were common to all site-specific models; however, the weight of these features differed. Conclusion A machine learning model for predicting hospital admission from the ED will generalize fairly well within the health care system but will still have significant differences in AUC performance across sites because of site-specific factors.
February 2022
·
55 Reads
·
4 Citations
Mayo Clinic Proceedings Innovations Quality & Outcomes
Objective To investigate medication factors and patient characteristics associated with readmissions following alcohol-related hospitalizations. Patients and Methods Adult patients admitted from September 1, 2016, through August 31, 2019, who had an alcohol-related hospitalization were identified through electronic health records. Patient characteristics and medications of interest administered during hospitalization or prescribed at discharge were identified. Medications of interest included US Food and Drug Administration–approved medications for alcohol use disorder, benzodiazepines, barbiturates, gabapentin, opioids, and muscle relaxants. The primary outcome was to identify medications and patient factors associated with 30-day alcohol-related readmission. Secondary outcomes included medications and patient characteristics associated with multiple alcohol-related readmissions within a year from the index admission (ie, two or more readmissions) and factors associated with 30-day all-cause readmission. Results Characteristics of the 932 patients included in this study associated with a 30-day alcohol-related readmission included younger age, severity of alcohol withdrawal, history of psychiatric disorder, marital status, and the number of prior alcohol-related admission in the previous year. Benzodiazepine or barbiturate use during hospitalization or upon discharge was associated with 30-day alcohol-related readmission (P=.006). Gabapentin administration during hospitalization or upon discharge was not associated with 30-day alcohol-related readmission (P=.079). Conclusion The findings reinforce current literature identifying patient-specific factors associated with 30-day readmissions. Gabapentin use was not associated with readmissions; however, there was an association with benzodiazepine/barbiturate use.
December 2021
·
53 Reads
·
8 Citations
December 2021
·
1,631 Reads
·
8 Citations
Mayo Clinic Proceedings Innovations Quality & Outcomes
October 2021
·
190 Reads
We introduce DeepABM, a framework for agent-based modeling that leverages geometric message passing of graph neural networks for simulating action and interactions over large agent populations. Using DeepABM allows scaling simulations to large agent populations in real-time and running them efficiently on GPU architectures. To demonstrate the effectiveness of DeepABM, we build DeepABM-COVID simulator to provide support for various non-pharmaceutical interventions (quarantine, exposure notification, vaccination, testing) for the COVID-19 pandemic, and can scale to populations of representative size in real-time on a GPU. Specifically, DeepABM-COVID can model 200 million interactions (over 100,000 agents across 180 time-steps) in 90 seconds, and is made available online to help researchers with modeling and analysis of various interventions. We explain various components of the framework and discuss results from one research study to evaluate the impact of delaying the second dose of the COVID-19 vaccine in collaboration with clinical and public health experts. While we simulate COVID-19 spread, the ideas introduced in the paper are generic and can be easily extend to other forms of agent-based simulations. Furthermore, while beyond scope of this document, DeepABM enables inverse agent-based simulations which can be used to learn physical parameters in the (micro) simulations using gradient-based optimization with large-scale real-world (macro) data. We are optimistic that the current work can have interesting implications for bringing ABM and AI communities closer.
June 2021
·
20 Reads
·
3 Citations
Journal of the American Medical Informatics Association
Hospital census prediction has well-described implications for efficient hospital resource utilization, and recent issues with hospital crowding due to CoVID-19 have emphasized the importance of this task. Our team has been leading an institutional effort to develop machine-learning models that can predict hospital census 12 hours into the future. We describe our efforts at developing accurate empirical models for this task. Ultimately, with limited resources and time, we were able to develop simple yet useful models for 12-hour census prediction and design a dashboard application to display this output to our hospital’s decision-makers. Specifically, we found that linear models with ElasticNet regularization performed well for this task with relative 95% error of +/− 3.4% and that this work could be completed in approximately 7 months.
... See code snippets in Figure 1 and Figure 2. AgentTorch's design is motivated by active collaborations around the world, with models currently being deployed to help mitigate a measles outbreak in New Zealand [9]; capture the foraging behavior of migratory birds in Alaska; and analyze the dairy supply chain in the Pacific islands to safeguard against a potential H5N1 outbreak. Previously, AgentTorch models were used across multiple countries during the COVID-19 pandemic to evaluate immunization protocols [42,13] deployed across multiple countries [24,34]. The project is under active development, and we refer to the GitHub repository for tutorial notebooks, videos, reference implementations, installation guides, and contribution guidelines. ...
October 2023
Vaccine
... Datasets sourced from EHRs offer rich longitudinal and pathophysiological data, which significantly bolster efforts in drug repurposing [115,116]. One notable study by Ryu et al. focused on monoclonal gammopathy of undetermined significance (MGUS), which is a non-cancerous hematological condition that may develop into malignant diseases such as multiple myeloma [118]. Leveraging machine learning (ML) and EHR data, the researchers analyzed information from a comprehensive MGUS database encompassing 16,752 patients diagnosed between 2000 and 2021 at the Mayo Clinic. ...
February 2023
Blood Cancer Journal
... The key properties of blockchain are decentralization, immutability, transparency, and distributed consensus. Blockchain enhances data security, traceability and authenticity that can safely store sensitive data like vaccination records and strengthen digital vaccination certificates [12,14]. A research study on the possibility of using blockchain and smart contract-based digital certificate was carried out [15]. ...
July 2022
Frontiers in Big Data
... Certain studies have revealed sources of bias, imprecision, and even increased time-load on physicians interacting with ML recommendation systems that have been deployed in EHR environments [14,15]. As applications of ML to EHR data become less hypothetical, current and future research efforts in this area will likewise need to become more practical, focusing on the challenges that inevitably arise as ML is integrated into real-time CDS [10,[16][17][18]. ...
June 2022
Mayo Clinic Proceedings Innovations Quality & Outcomes
... Repast4Py (Python) [22] aims to lower the entry barrier for modelers to try distributed simulators by simplifying the process. DeepABM [23] stores each agent and their states in tensors and models the interactions between them using graph neural networks (GNNs). ...
December 2021
... Cloud computing and virtualisation, however, are rapidly becoming the norm in the health sector. [26][27][28] These advances have given rise to digital research environments, which permit secure, auditable data sharing and analysis within a virtual workspace, 29 eliminating the need to transfer data between local hard drives. In the Netherlands, the AnDREa project is one of the numerous working examples of such endeavours, 29 and PRISMA data will be accessible within this environment. ...
December 2021
Mayo Clinic Proceedings Innovations Quality & Outcomes
... [13][14][15][16][17][18] One study used linear regression to predict future bed occupancy based on the current and past hospital bed occupancy. 19 While some of these studies demonstrated good predictive accuracy, the statistical knowledge and experience required to apply and interpret these models are often absent in local health systems. 20 In North West London (NWL), the National Health Service (NHS) response to the pandemic was coordinated and strategically led by the COVID-19 NWL Gold Command group, where senior representatives from across the integrated care system would come together at least once a week to look at operational and strategic actions, which were required to support the system. ...
June 2021
Journal of the American Medical Informatics Association
... By leveraging DT platforms, researchers can gather large data more quickly and from a broader population base. Simulations and virtual environments offer a unique opportunity for policymakers to test the implications of healthcare policies in a controlled, risk-free setting [40]. By modeling the outcomes of proposed policies, stakeholders can anticipate their effects and refine them before implementation, ensuring that new regulations are both effective and efficient. ...
May 2021
The BMJ
... In fields such as food processing, medical diagnosis, and public safety, alcohol is a commonly encountered chemical substance [10,11]. The measurement of alcohol concentration is crucial for ensuring people's health and safety [12]. ...
April 2021
Mayo Clinic Proceedings Innovations Quality & Outcomes
... values, beliefs, ethical principles and morality), and provide support, counselling and education plus various rituals and rites of passage in order to ensure holistic care" (Carey & Rumbold, 2015, p. 1222. In a systematic review of observational and experimental studies regarding the use of chaplains in the hospital setting, spiritual care was correlated with increased patients' satisfaction (Kirchoff et al., 2021). The care given by chaplains was also associated with spiritual well-being and better quality of life (Kirchoff et al., 2021). ...
June 2021
Journal of Religion and Health